This message was deleted.
# harvester
a
This message was deleted.
same applies to a dual-node setup.
if you dont want the degraded warning you must set up a new storage class. otherwise it works fine.
g
cool and thanks, so I need to remove standard storage-class and create new? can I tweak original to be another replicate number?
c
iirc the default ones are fixed. but you can simply create a new one with a replica set to 2 (for your case), and sort your new VM volumes under that new class.
you can also update the default storage class to be used to your new one.
g
thanks, so can I remove old one when I select new as default?
c
I cant say anything about that but since it is only a storage class, you can simply keep it, as long as your volumes are created under your new storage class, nothing will change for you. I dont quite know what else is running under those 2 default storage classes.
g
I have 4 disks in each server, should I create a RAID 5 for those or keep them as single disks and let longhorn handle data?
c
well, longhorn with replica set to 2 effectily is a network raid (raid 1)
it is a block storage.
disks are added to longhorn through the harvester dashboard on an individual basis, unformatted and unpartioned. longhorn does the rest 🙂
g
yes I know, just if I should have 4x1.8TB or 1x5.4TB (hardware RAID) in each server - I guess if I don’t need a vm data disk higher than 1.8 4x1.8TB is maybe better
c
ooh hardware raid, not software raid. I see. keep them separate I think. raid 0 always gives me creeps 😛
also longhorn doesnt like huge data disks.
just keep that in mind
g
cool will try JBOD mode and give each server 4 single disks
👍 2
☝️ 1
besides OS disks
b
> I have 4 disks in each server, should I create a RAID 5 for those or keep them as single disks and let longhorn handle data? (edited) Second +1 for leaving them as single disks and letting longhorn deal with it.
1